November 13, 2015 - by David Hess
Our TeamRankings Football Championship regular season concludes this week. There are three remaining spots in the $25,000 Week 11 Final up for grabs, and they’ll be awarded to the top three scores in this week’s contest.
And as the TRFC contest comes to an end, so do my weekly FanDuel NFL lineup posts.
After a successful foray playing NFL on FanDuel for the first time in 2014, I had high hopes for 2015.
We haven’t yet built any daily fantasy projection models or analytical tools here at TeamRankings, so I knew my DFS competitiveness would be handicapped to some degree.
Still, I felt like as long as I could get my hands on some decent player projections and optimization tools, common sense combined with my knowledge of game theory principles would give me a decent edge.
Last year’s success (doubling a $250 bankroll in 8 weeks) also provided encouragement.
However, in 2015 I’ve now lost over half my bankroll in nine weeks, and recent results have not been getting any better.
If I’ve learned anything this year, it’s that my general approach last season — whipping up some lineups on Friday morning, and basing my player selections mostly on widely available weekly projections — has not been a winning strategy.
Maybe it never was, and I just got lucky last year. It’s tough to tell given the small sample size of my results to date.
In retrospect, I suppose this shouldn’t be all that surprising. In our pick’em and survivor pool Q&A forums, we often point out that betting markets are very efficient, and it’s difficult to gain a big edge if you’re basing your pick decisions on widely-known information.
For example, I’ve been using numberFire’s player projections as one of my data sources this year. numberFire was already pretty well known in the fantasy community, but now that the site’s been acquired (and heavily promoted) by FanDuel itself, I’ve probably been sharing a core source of research with thousands of people who can compete against me on FanDuel.
In addition, daily fantasy sports have exploded in popularity this season, so it stands to reason that the market may have gotten more efficient compared to last year.
Sure, there are more “squares” or “fish” (or whatever you want to call dumb money) playing DFS now. But it’s pretty much a given that many more big-money pros and syndicates are operating as well, and the structure of DFS lets them enter tons of contests with ease.
If I make another foray into daily fantasy in the future, it seems that leveraging our historical database and predictive modeling skills here at TeamRankings ought to be part of the plan.
Creating lineups based on our own internal player projections could be fruitful; at least then our lineup choices would probably be more unique, and more likely to pay off in weeks when they have high accuracy and competing projection systems don’t perform quite as well.
However, it’s a big undertaking to create an accurate player projection system, and we’d have to prioritize it against other opportunities.
There are also several analytical tools for DFS that we’ve discussed developing, that don’t seem to exist yet. But if we built them, we’d have to decide whether it would be better to share or sell them on the site, or just to use them to play ourselves, presumably for higher stakes.
In the end, losing sucks, but it’s still been a fun season. I continued to learn a lot about DFS, and we’ll see if I can salvage a winning year with some huge victory as my final hurrah this week. 🙂
Whatever happens this Sunday, I’ll likely continue to play in my own free time, at lower stakes, just so I’m more emotionally invested in the games.
This week I went about my lineup selection a bit differently than in past weeks.
Instead of first using a very focused head-to-head and 50/50 strategy and then moving on to create tournament lineups separately, I created tournament lineups first and then used the players that ended up in multiple tournament lineups in my primary 50/50 lineup.
My reasoning for this was that I’ve noticed in several past weeks that I’ve had multiple good value player selections filling in around my stacks in my tournament lineups. But I rarely had all those value picks in a single lineup, so even though several of them did well, I never got the benefit of their combined positive performances.
Since my previous 50/50 strategy didn’t really seem to be working, I figured switching it up this final week couldn’t hurt.
Here’s my primary lineup for Week 10:
Some notes on this lineup:
Here’s my secondary lineup:
Notes on this lineup:
As is the case most weeks, my first step this week was to come up with some single-team stacks that seem to have the potential for a big day.
Again this week I focused on something pretty basic — picking teams that are projected to score a lot of points. And by “focused on,” I mean I simply picked the 5 teams with the highest projected scores according to the betting markets.
As for the selections of which players to stack, those were almost entirely driven by looking at Football Outsiders’ data on passing yards allowed to different types of receivers. I took players who play positions the opponents have been poor against this season, where possible.
Here’s what I ended up with:
Filling in around the stacks with solid consensus values, with some emphasis on trying to find players with high upside, here’s the result (click to enlarge):
As usual, I’ll be entering these lineups in a FanDuel Sunday Million tournament, as well as a smaller 100-person or 250-person tournament. The idea is that if they do astronomically well, I’ll be rewarded for that in the Sunday Million, but if they just score in the top 1% or so, I’ll get a better return in the 100-person tournament.
The most likely outcome for any specific tournament lineup is that I win nothing, but with multiple stacks, hopefully at least one will still do well enough to recoup my buy-ins.
Printed from TeamRankings.com - © 2005-2019 Team Rankings, LLC. All Rights Reserved.